“…Seven nonrandomized studies with one arm's sample size being at least twice as big as the other arm and which had high chances of bias were excluded. 11,12,14,25,27,28,31 The 24 remaining studies were analyzed, yielding similar results: LA was associated with shorter total surgical time (WMD À8.77 min [À16.84 to À0.69]; p = 0.03) ( Supplementary Fig 3), lower stroke rate (OR 0.64 [0.38-0.89]; p = 0.01) ( Supplementary Fig 4), cardiac complications (OR 0.48 [0.32-0.72]; p = 0.004) ( Supplementary Fig 5), 30-day mortality (OR 0.66 [0.45-0.96]; p = 0.03) ( Supplementary Fig 6), and TND rates (OR 0.58 [0.38-0.89]; p = 0.01) ( Supplementary Fig 7). Heterogeneity was significant for total surgical time (I 2 = 0.99, chisquare = 1,214.45; p < 0.00001) and cardiac complications (I 2 = 0.43, chi-square = 29.79; p = 0.03) but not for stroke (I 2 = 0.15, chi-square = 18.88; p = 0.28), TND (I 2 = 0.24, chisquare = 18.33; p = 0.19), and 30-day mortality (I 2 = 0.00, chisquare = 11.72; p = 0.86).…”